*** Instructions ***

This replication capsule runs on Code Ocean, which uses an AWS EC2 instance with 16 cores and 120 GB of memory. The complete replication runs in about 1-2 hours.

Once complete, publication figures and tables will appear in ./results/for_publication/figures/ and ./results/for_publication/tables/, respectively. All other files in the ./results/ directory are intermediate data, figures, and tables.

To intitiate the replication process in another environment, set the working directory to the ./code/ folder and then run the bash script run.sh.

*** Content Notes ***

* ./data/ *

- politicaldomains-final-edits.csv: list of news domains with categories from Peterson et al. (2019)

- news_domain_urls.csv: timestamped URL visits to news domains.. this file is derived from the original dataset by dropping URL visits to non-news domains (given in politicaldomains-final-edits.csv)

- data_for_usage_tables.csv: for each caseid (i.e., respondent) x date pair, counts the total number/duration of all URL visits, facebook visits, news visits, etc.

- weights.csv: the survey weight for each caseid (i.e., respondent); also available in the original survey data files but stored in this csv file for easy access

- google_trends_clinton_trump.csv: Google trends data

- PoliticalKeyWords.csv: the list of political keywords described in the main text.

- top500.csv: the websites and associated scores from Bakshy et al.

- bazaars_figure_map.csv: a spreadsheet that maps the intermediate figure references onto the their publication references (e.g., main_figure_3.pdf)

- bazaars_table_map: a spreadsheet that maps the intermediate table references onto the their publication references (e.g., main_table_3.tex)

- November Survey/: data and codebook for the survey component of the web browsing data 

* ./code/ *

- Note that there is no 1_xyz.R file included in this replication capsule, since that file is used to refine the (much larger) raw data files into the processed files provided in /data/.

* ./results/ *

NOTE: Appendix Table 1 is a list of chosen domains and thus does not need to be replicated. Additionally, we cannot share the data to generate Appendix Table 4, since it is part of a separate data collection and associated projects. All other tables and figures should be generated by successfully executing the replication code.

- Numerical results and any temporary files will go here along with table and figure subfolders.

- Publication figures and tables will appear in ./results/for_publication/figures/ and ./results/for_publication/tables/, respectively.